A/B testing forms: what to test and how to test it

You set up a form. You wrote what you thought was clear copy. You designed it to look clean and professional. Then you checked the submissions and wondered why so few people finished it.

The assumption most brands make is that their form is fine. The real answer might be that a single change would double your submission rate. A different button color. Shorter copy on one field. Removing one optional question. One small change could be the difference between a form that converts and one that sits there collecting dust.

That is what A/B testing forms does. It replaces guessing with data. You test two versions of a form against real visitors and measure which one gets more submissions. The winning version becomes your new baseline. Then you test again. Over time, small improvements compound into dramatically better conversion rates.

This chapter covers what to test, how to run form experiments, what results tell you, and how to use testing to build a form that converts instead of one that just exists on your website.

What makes a form element worth testing

Not every part of a form deserves a test. If you spend time testing things that do not matter, you will waste your testing budget without learning anything useful.

The elements worth testing are ones that directly influence whether someone submits the form or abandons it. These fall into a few categories.

High-friction elements that might cause abandonment

Friction is anything that makes the visitor hesitate, confused, or reluctant. A form with too many fields has high friction. One that asks for information before the visitor has context for why they need to provide it has friction. Permission requests that sound spammy have friction.

Removing friction almost always improves conversion. If studies show 70% of online shoppers abandon carts when asked for unnecessary information, what do you think happens to forms with the same problem? Test anything that feels like it might be turning visitors away.

Elements that directly ask for commitment

Call-to-action buttons, permission requests, and form headlines are where visitors make the actual decision to commit. These deserve testing because small changes in wording can shift how the visitor perceives what they are about to do.

A button that says "Send my request" feels different than one that says "Start my free trial." Same action, different perception of commitment. That difference matters.

Elements that build trust or create doubt

Privacy reassurance, security badges, guarantees, or money-back promises all influence whether the visitor trusts you enough to fill out the form. Permission copy, data collection requests, and terms language either build trust or undermine it. These are worth testing because they directly affect conversion.

An ecommerce checkout asking for phone number with no explanation will have lower conversion than the same form with a line of copy that says "We use this for shipping and delivery updates only." That small addition might lift conversion by 10% or more.

The five form elements most worth testing first

If you are new to form testing, start with the elements that historically move conversion rates the most. Research across thousands of form tests consistently shows that certain elements have higher impact than others.

Number of form fields

This is the single most-tested element in form optimization because it has the biggest impact on conversion. Every field you add to a form increases abandonment. Every field you remove increases completion.

The challenge is that more fields mean more data collection. If you need 20 fields of information from the visitor, you cannot just remove 19 and expect better conversions. But if you ask for 20 pieces of information you do not need, then yes, removing those fields will dramatically improve conversion.

A real example: a SaaS company tested removing optional fields from their signup form. They went from asking for 20 fields to asking for just four. First name, email, company, and password. Everything else was collected after signup. The result was a 188% increase in completed signups.

But here is the trade-off. With fewer fields, they also had fewer qualified leads. Not every signup was a sales-ready prospect. They were able to sort quality by engagement after signup instead of trying to guess at signup time. This works if your business model can handle unqualified leads. It does not work if you need pre-qualified information upfront.

When testing field count, be strategic about which fields are truly required and which are nice-to-have. Test removing the nice-to-have ones first.

Call-to-action button copy

The button text is the last thing the visitor reads before committing. Wording matters. A button that says "Send" feels passive. One that says "Get my free guide" feels active and outcome-focused.

One of the most famous button tests in e-commerce changed "Start your free 30 day trial" to "Start my free 30 day trial." The shift from impersonal to personal language increased click-through by 90%. That single word changed how visitors perceived the action.

For contact forms, test "Send message" vs "Schedule a call." For newsletter signups, test "Join our list" vs "Get weekly tips." For checkout, test "Complete order" vs "Get my order today." Each variation reflects a different perception of what happens next. One will perform better with your specific audience.

Form structure: single-step vs. multi-step

Single-page forms are simpler. Multi-step forms (also called progressive disclosure) ask one question at a time or group related fields on separate pages. Studies show that multi-step forms often increase completion because each step feels smaller and less overwhelming.

However, the test results depend on your form length. A 4-field form does not need multiple steps. A 15-field form probably does. A donation form tested splitting payments onto a separate page and increased completion by 26.5% because the form stopped looking so long.

The win here is psychological. Each step you move to a new page signals progress. The form stops feeling endless. Visitors feel like they are getting closer to the finish.

Privacy and permission copy

Permission language either encourages submission or discourages it. Forms that ask for permission (email signups, newsletter opts, data tracking consent) see conversion lifts when the copy becomes more specific and trust-focused.

A generic checkbox that says "Subscribe to our newsletter" might have a 30% opt-in rate. The same form with a checkbox that says "Get weekly deals on items in my favorite categories. One email per week. Unsubscribe anytime" might have a 45% opt-in rate. Specificity and reassurance change behavior.

Test different permission language. Be more specific about frequency and benefit. Add reassurance about data use or unsubscribe ease. Measure which version gets more opt-ins.

Form field labels and help text

Labels tell the visitor what information to put in each field. Help text (the smaller copy below or beside the field) clarifies why you need that information.

Confusing labels cause errors. Extra errors cause abandonment. A field labeled "Company" does not explain whether you mean the company the visitor works for or owns. Adding help text like "Enter the name of your company or organization" removes the guesswork.

Test clearer labels, better help text, or different label placement. Inline validation (highlighting fields as the visitor fills them) can also improve completion by giving immediate feedback instead of showing all errors at the end.

How to structure an A/B test for forms

Running a form test is straightforward, but doing it correctly matters. A poorly designed test will give you false confidence in results that are unreliable.

Step 1: Start with a hypothesis

Do not just guess. Hypothesis-driven testing gives you a framework for thinking about what might move the needle.

Your hypothesis should follow this structure: "If I [change this element], then [this metric will improve], because [this is the reason I think it will work]."

For example: "If I change the button copy from 'Send request' to 'Schedule my demo,' then submission rate will increase by at least 10%, because the new copy makes it clearer what action they are taking and reduces psychological friction."

The reason matters because it forces you to think about whether the change addresses a real problem or if you are just guessing.

Step 2: Run the test to enough visitors

A form test needs statistical significance to be reliable. If your form gets 100 visitors per month, running a test that only exposes 10 visitors to the new version will not give you reliable data.

Most form tools will tell you when a test has reached statistical significance. Generally, you need at least 100 conversions in each variant to trust the results. If your form gets only a handful of submissions per week, you will need to run the test for weeks or months to get reliable data.

The test should run for at least one full week to avoid day-of-week effects (maybe fewer people submit forms on Fridays). Ideally, run for two weeks to a month so seasonal patterns do not skew results.

Step 3: Measure the right metrics

The metric that matters is the conversion rate of the form. Conversion rate is the percentage of visitors who see the form and complete it.

This is different from submission rate. A visitor who sees the form and does not fill it out is part of the conversion calculation. That is why form tests often move conversion even when the number of submissions is relatively low.

Secondary metrics also matter. If you are testing a change that might affect lead quality, track quality metrics too. A test that increases signups by 50% but decreases average deal size by 75% is a loss, not a win.

Step 4: Only change one element at a time

If you change the button copy, the number of fields, and the help text all at once, you will not know which change moved the needle. A/B testing isolates variables. One change per test. One winning version becomes the new baseline. Then test the next element.

This takes longer than making five changes at once, but it tells you what works instead of what you guessed might work.

Step 5: Implement the winning version

Once a test wins, implement the winning version permanently. Set it as your control. This becomes your new starting point for the next test.

Do not go back to the original version just because it was the original. The test results told you the new version is better. Stick with it unless a future test beats it.

What results tell you

Form test results show what works with your specific audience. The key word is your. A button color that increased conversion by 20% on someone else's form might do nothing on yours. Your audience is different. Your form context is different. Your offer is different.

Use test results to learn about your audience, not to apply universal rules. Some audiences respond to urgency language. Others find it manipulative. Your tests will show you which approach works for people interested in your specific offer.

Also understand that small improvements compound. A 5% improvement on a form might sound tiny. But if your form generates 1,000 submissions per month, a 5% improvement is 50 more submissions. Over a year, that is 600 more submissions from the same traffic.

Run three months of regular testing, implement three winning variants, and you might see a cumulative 15-20% improvement in conversion rate. That compounds fast.

Common testing mistakes to avoid

Several mistakes will make your test results unreliable or misleading.

Running tests for too short a time period

If you run a test for two days and then declare a winner, you are probably wrong. You need at least 100 conversions per variant to have reliable data. You also need to account for day-of-week effects and variability in traffic patterns. Run tests for at least one week, preferably two to four weeks.

Changing multiple elements in one test

If you test both a new button color and new button copy at the same time, you will not know which one moved the needle. You might even have one change that improved conversion and another that hurt it, and they offset each other. You will not see the net effect.

Testing things that will not matter

Spend testing time on high-impact elements. Testing whether your form should have a white background or off-white background is not worth your time. Testing whether removing two optional fields increases completion rates is worth it.

Stopping the test too early

Do not watch the test results day-by-day and stop as soon as one variant looks like it is winning. Tests need time to stabilize. A test that looks like a winner on day three might reverse by day ten. Let the test run to your predetermined end date, then declare the winner based on final results.

Ignoring secondary metrics

If your test increases form submissions but decreases the quality of leads submitted, the test failed. Track not just volume but the metrics that matter to your business. For leads, track sales conversion. For newsletter signups, track email engagement. For checkout, track average order value. One metric in isolation does not tell the whole story.

A/B testing forms with WEMASY

WEMASY's form builder lets you run A/B tests without touching code. You can test different form versions and let the system automatically split traffic between them.

Use form analytics to track which variant performs better. WEMASY shows you submission rates, field-by-field abandonment, and conversion funnels so you can see where visitors are dropping off and which version of your form is winning.

Start with testing high-impact elements. Remove optional fields and measure submission rate improvement. Test different call-to-action button copy and measure click-through rates. Test whether your form needs to be multi-step. Let your visitors tell you what works.

For deeper strategy on form optimization, learn about reducing form abandonment and writing form copy that converts. Explore WEMASY's form builder and start testing your forms against real visitors. See everything about forms for the complete picture of form strategy.

Frequently asked questions